INDEX

Explanations

negative words or terms related to unfavorable treatment

oai_token-act-pair · gpt-3.5-turbo

references to various types of treatment and perceptions regarding individuals or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 impossibility

-0.84

igsaw

-0.79

UNCH

-0.73

ausible

-0.73

ospel

-0.69

rame

-0.64

 impro

-0.62

rossover

-0.61

aghetti

-0.60

POSITIVE LOGITS

 toward

1.15

 towards

1.12

itism

0.93

 Towards

0.89

 demeanor

0.86

 greeted

0.83

nered

0.81

 attitude

0.79

eus

0.79

Semitic

0.78

Activations Density 0.330%

negative words or terms related to unfavorable treatment

references to various types of treatment and perceptions regarding individuals or groups

No Comments

No Known Activations