INDEX

Explanations

names of people or entities being mentioned in a conversation or text

oai_token-act-pair · gpt-3.5-turbo

references to political statements or decisions

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

grown

-0.78

 rotting

-0.75

 Warehouse

-0.70

 goddamn

-0.69

 sewer

-0.69

 sewage

-0.69

igslist

-0.68

itary

-0.68

 Craigslist

-0.68

 civilization

-0.67

POSITIVE LOGITS

 rhet

0.96

 sarcast

0.91

 emph

0.90

 reiter

0.87

Asked

0.85

kish

0.84

Comment

0.80

 emphas

0.80

answered

0.79

 Asked

0.79

Activations Density 0.809%

names of people or entities being mentioned in a conversation or text

references to political statements or decisions

No Comments

No Known Activations