INDEX

Explanations

keywords related to legal and official documents or processes

oai_token-act-pair · gpt-3.5-turbo

references to formal documents or reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

rocket

-0.57

à

-0.55

 brill

-0.53

conservancy

-0.52

 skiing

-0.52

 Xiaomi

-0.50

 riot

-0.50

 rocket

-0.50

 Trails

-0.49

wagen

-0.48

POSITIVE LOGITS

soType

0.62

 concludes

0.61

ENTS

0.56

enum

0.55

KE

0.55

 advises

0.55

 quotes

0.54

 explained

0.54

 GOODMAN

0.54

sonian

0.54

Activations Density 1.144%

keywords related to legal and official documents or processes

references to formal documents or reports

No Comments

No Known Activations