INDEX

Explanations

words related to negative effects or connotations

oai_token-act-pair · gpt-3.5-turbo

references to negative experiences or impacts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nington

-0.90

plain

-0.87

DOM

-0.84

aukee

-0.80

frey

-0.79

ISTORY

-0.79

aido

-0.78

ingham

-0.77

xxxxxxxx

-0.77

keeping

-0.77

POSITIVE LOGITS

 consequences

1.04

 impact

0.98

 reinforcement

0.97

 impacts

0.96

 repercussions

0.96

 feedback

0.96

 gearing

0.96

 consequence

0.95

 publicity

0.95

 reperc

0.93

Activations Density 0.045%

words related to negative effects or connotations

references to negative experiences or impacts

No Comments

No Known Activations