INDEX

Explanations

mentions or instances of the word "ill"

oai_token-act-pair · gpt-3.5-turbo

the word "ill" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

oÄŁ

-0.78

 critically

-0.64

İĭ

-0.64

 Aval

-0.63

blance

-0.62

 compr

-0.62

 slack

-0.62

é¾įå¥ĳå£«

-0.62

EStream

-0.61

¥µ

-0.60

POSITIVE LOGITS

espie

1.35

uminati

1.25

icit

1.05

inois

1.01

inger

0.97

iman

0.96

iard

0.92

igan

0.91

ieri

0.88

oyd

0.87

Activations Density 0.009%

mentions or instances of the word "ill"

the word "ill" in various contexts

No Comments

No Known Activations