INDEX

Explanations

terms indicating a change in name or previous identity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "formerly" indicating past identities or names

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anche

-0.97

eer

-0.96

otle

-0.90

antics

-0.85

eers

-0.84

oos

-0.81

pled

-0.81

atics

-0.76

rer

-0.75

uri

-0.74

POSITIVE LOGITS

 obsolete

0.94

 incarcerated

0.86

 exting

0.85

 dubbed

0.84

 housed

0.84

 known

0.82

etheless

0.82

 defunct

0.82

 unsus

0.80

 redes

0.77

Activations Density 0.015%

terms indicating a change in name or previous identity

instances of the word "formerly" indicating past identities or names

No Comments

No Known Activations