INDEX

Explanations

words related to criticism or negative evaluation

oai_token-act-pair · gpt-3.5-turbo

terms related to significant consequences or effects

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

 Awakens

-0.68

 Kardash

-0.66

 Masquerade

-0.57

 fitt

-0.54

 trusts

-0.52

 didnt

-0.51

 Kenn

-0.51

 retrie

-0.50

Wes

-0.49

Mak

-0.49

POSITIVE LOGITS

lie

0.72

maxwell

0.68

ieu

0.68

JECT

0.62

onite

0.61

olina

0.61

ril

0.61

oton

0.59

acus

0.58

inguished

0.58

Activations Density 1.571%

words related to criticism or negative evaluation

terms related to significant consequences or effects

No Comments

No Known Activations