INDEX

Explanations

references to figures or specific sections within a document

oai_token-act-pair · gpt-3.5-turbo

references to figures or diagrams

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

pron

-0.78

izabeth

-0.75

acia

-0.72

umenthal

-0.70

eworld

-0.69

ances

-0.68

resist

-0.68

nee

-0.68

 homeland

-0.67

itch

-0.65

POSITIVE LOGITS

Figure

1.05

 Figure

0.86

Ī

0.80

book

0.75

 Explorer

0.75

Fig

0.74

figure

0.74

 Drawing

0.72

Table

0.71

 Sheet

0.71

Activations Density 0.019%

references to figures or specific sections within a document

references to figures or diagrams

No Comments

No Known Activations