INDEX

Explanations

expressions related to negative judgment or criticism

oai_token-act-pair · gpt-3.5-turbo

derogatory remarks about intelligence

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

AUT

-0.92

 largeDownload

-0.88

APH

-0.87

quart

-0.76

cussion

-0.69

HI

-0.69

aver

-0.67

RH

-0.67

ILA

-0.67

soType

-0.66

POSITIVE LOGITS

 stupid

1.01

nesses

0.93

 silly

0.85

 Stupid

0.84

 dumb

0.83

gery

0.79

itude

0.77

upid

0.77

ulent

0.77

ishly

0.76

Activations Density 0.010%

expressions related to negative judgment or criticism

derogatory remarks about intelligence

No Comments

No Known Activations