INDEX

Explanations

phrases or words indicating disapproval or lack of acceptance

oai_token-act-pair · gpt-3.5-turbo

statements expressing condemnation or disapproval of actions deemed inappropriate or immoral

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ynthesis

-0.88

zyme

-0.85

tein

-0.84

pel

-0.79

oult

-0.77

ingers

-0.76

uay

-0.76

ools

-0.76

hung

-0.75

ulet

-0.74

POSITIVE LOGITS

 compromises

0.84

 adolesc

0.82

 unacceptable

0.76

 behaviour

0.73

 occurrences

0.72

 srfAttach

0.71

 undermin

0.70

 ambassadors

0.69

 enrichment

0.68

LY

0.68

Activations Density 0.015%

phrases or words indicating disapproval or lack of acceptance

statements expressing condemnation or disapproval of actions deemed inappropriate or immoral

No Comments

No Known Activations