INDEX

Explanations

verbs associated with personal actions or decisions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating emotional or relational connections

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ELS

-0.56

Eastern

-0.53

hawks

-0.52

los

-0.52

 Adin

-0.52

 Loaded

-0.50

Bridge

-0.50

Il

-0.50

Els

-0.49

thodox

-0.49

POSITIVE LOGITS

 yourself

1.22

 yourselves

1.11

 Yourself

0.88

 your

0.78

 YOUR

0.71

your

0.66

Your

0.65

 Your

0.61

poke

0.60

 browsing

0.55

Activations Density 0.715%

verbs associated with personal actions or decisions

phrases indicating emotional or relational connections

No Comments

No Known Activations