INDEX

Explanations

information related to locations, specifically in the U.S

oai_token-act-pair · gpt-3.5-turbo

periods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 mosqu

-0.64

 advis

-0.59

 transferring

-0.56

 ambush

-0.56

 summ

-0.55

forth

-0.54

 onboard

-0.54

ucci

-0.54

 scaling

-0.53

pus

-0.53

POSITIVE LOGITS

1.27

$.

1.12

 States

0.96

Va

0.95

0.90

States

0.87

Ds

0.85

 STATES

0.84

Ns

0.83

Activations Density 0.043%

information related to locations, specifically in the U.S

periods

No Comments

No Known Activations