INDEX

Explanations

words related to specific entities or titles likely related to a brand or product called "Monster"

oai_token-act-pair · gpt-3.5-turbo

words and patterns related to specific fictional characters or titles, particularly those associated with "Mon" or "Mono."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BW

-0.64

ETF

-0.61

NER

-0.59

 stewards

-0.58

utenberg

-0.57

ONSORED

-0.56

enegger

-0.56

 hurd

-0.56

 divest

-0.55

negie

-0.55

POSITIVE LOGITS

ciating

0.77

liction

0.74

tick

0.73

atari

0.70

oxide

0.70

etheus

0.68

imester

0.68

achine

0.67

jong

0.66

nant

0.65

Activations Density 0.067%

words related to specific entities or titles likely related to a brand or product called "Monster"

words and patterns related to specific fictional characters or titles, particularly those associated with "Mon" or "Mono."

No Comments

No Known Activations