INDEX

Explanations

expressions of strong emotions or opinions

oai_token-act-pair · gpt-3.5-turbo

the word "totally" and its variations, indicating strong emphasis or agreement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

rers

-0.84

åº

-0.76

ently

-0.75

lings

-0.73

sburg

-0.73

mere

-0.72

issance

-0.71

pring

-0.70

llular

-0.69

llor

-0.69

POSITIVE LOGITS

 understandable

0.75

 unrelated

0.71

 forget

0.68

 freaking

0.68

 obliter

0.67

 annihil

0.65

 surprised

0.65

 allergic

0.65

MIA

0.64

 unexpected

0.64

Activations Density 0.021%

expressions of strong emotions or opinions

the word "totally" and its variations, indicating strong emphasis or agreement

No Comments

No Known Activations