INDEX

Explanations

phrases expressing opinions or attributions towards people or groups

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ondo

-0.79

inx

-0.71

urgical

-0.69

urgy

-0.66

 externalToEVAOnly

-0.65

amar

-0.64

iop

-0.64

abba

-0.63

 Supports

-0.63

issues

-0.63

POSITIVE LOGITS

 strangers

0.97

 comprised

0.93

 wealthier

0.93

 outnumbered

0.92

 minorities

0.91

 Caucas

0.91

 invariably

0.90

 overwhelmingly

0.90

 Asians

0.89

 legion

0.88

Activations Density 0.289%

phrases expressing opinions or attributions towards people or groups

No Comments

No Known Activations