INDEX

Explanations

names of prominent public figures

oai_token-act-pair · gpt-3.5-turbo

sequences of letters that appear to form parts of words or names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

staking

-0.62

âĶĢâĶĢ

-0.52

 Reincarn

-0.50

 Intelligent

-0.50

âĢ¢âĢ¢

-0.49

 anchors

-0.49

ĸļ

-0.49

shapeshifter

-0.48

 glasses

-0.48

REDACTED

-0.48

POSITIVE LOGITS

inski

0.84

encia

0.82

hett

0.78

ov

0.77

insky

0.74

berg

0.74

acus

0.72

yna

0.72

inav

0.72

nikov

0.71

Activations Density 0.131%

names of prominent public figures

sequences of letters that appear to form parts of words or names

No Comments

No Known Activations