INDEX

Explanations

requests for assistance or help

oai_token-act-pair · gpt-3.5-turbo

references to the act of helping or assisting others

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

é¾

-0.75

ithe

-0.67

ebook

-0.66

Pict

-0.65

gran

-0.65

aval

-0.64

ols

-0.64

punk

-0.64

War

-0.63

 Barg

-0.60

POSITIVE LOGITS

fully

0.88

 facilitate

0.84

 help

0.82

meet

0.81

 broker

0.77

aid

0.77

 Assist

0.74

 alleviate

0.73

tackle

0.71

help

0.71

Activations Density 0.039%

requests for assistance or help

references to the act of helping or assisting others

No Comments

No Known Activations