INDEX

Explanations

phrases related to negative or critical opinions about something

oai_token-act-pair · gpt-3.5-turbo

negative expressions or sentiments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Shank

-0.66

 strengthened

-0.65

 Reloaded

-0.64

Tik

-0.63

 hardness

-0.59

 Irwin

-0.59

DERR

-0.59

untled

-0.59

 Nicarag

-0.59

 hardened

-0.58

POSITIVE LOGITS

recomm

0.96

mom

0.93

distance

0.93

years

0.92

favorite

0.91

diff

0.91

dri

0.90

prison

0.90

dist

0.89

comments

0.88

Activations Density 0.109%

phrases related to negative or critical opinions about something

negative expressions or sentiments

No Comments

No Known Activations