INDEX

Explanations

descriptions and occurrences of specific legal or criminal actions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

contextual words related to health and safety in various situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 hemor

-0.76

 pestic

-0.72

 Canaver

-0.68

displayText

-0.65

 conduc

-0.64

 warr

-0.61

 caut

-0.60

 resil

-0.60

pse

-0.59

 reper

-0.59

POSITIVE LOGITS

their

0.68

obe

0.64

0.63

enders

0.60

ousands

0.58

 their

0.57

ers

0.56

bara

0.56

its

0.56

ants

0.56

Activations Density 0.599%

descriptions and occurrences of specific legal or criminal actions

contextual words related to health and safety in various situations

No Comments

No Known Activations