INDEX

Explanations

names of individuals in executive positions and titles of their organizations

oai_token-act-pair · gpt-3.5-turbo

names and titles of individuals associated with organizations or positions of authority

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

 chilling

-0.70

 symptoms

-0.68

 symptom

-0.67

 effect

-0.63

kinson

-0.63

 exact

-0.60

 hump

-0.60

 receipt

-0.60

 Russo

-0.60

 specifics

-0.59

POSITIVE LOGITS

 fame

0.88

å§«

0.72

responsible

0.71

eworks

0.65

bg

0.65

iatrics

0.65

 Lebanon

0.63

Ow

0.63

ignty

0.62

gem

0.62

Activations Density 0.597%

names of individuals in executive positions and titles of their organizations

names and titles of individuals associated with organizations or positions of authority

No Comments

No Known Activations