INDEX

Explanations

words related to technology brands or products

oai_token-act-pair · gpt-3.5-turbo

mentions of specific brands or proper nouns related to popular culture

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

brance

-0.71

aturdays

-0.66

ufact

-0.65

bender

-0.64

PDATE

-0.64

 compr

-0.61

 afore

-0.60

heit

-0.59

swick

-0.59

BG

-0.59

POSITIVE LOGITS

reme

0.97

Large

0.81

thur

0.78

abytes

0.74

©¶æ

0.73

atche

0.71

isoft

0.71

onential

0.71

aston

0.68

xes

0.67

Activations Density 0.124%

words related to technology brands or products

mentions of specific brands or proper nouns related to popular culture

No Comments

No Known Activations