INDEX

Explanations

descriptions of actions or situations that are problematic or controversial

oai_token-act-pair · gpt-3.5-turbo

words related to faintness or subtlety

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 restruct

-0.75

 hairc

-0.68

ecycle

-0.66

DW

-0.66

 haircut

-0.64

Delivery

-0.63

 Corsair

-0.62

ihad

-0.62

eneg

-0.60

 conversion

-0.59

POSITIVE LOGITS

 faint

3.91

ply

1.54

 slightest

1.42

 faintly

1.20

 sque

1.08

 indist

1.06

 blush

1.02

 feeble

1.00

 cries

0.95

 frail

0.86

Activations Density 0.023%

descriptions of actions or situations that are problematic or controversial

words related to faintness or subtlety

No Comments

No Known Activations