INDEX

Explanations

references to spatial locations or events

oai_token-act-pair · gpt-3.5-turbo

references to actions involving shutting or closing something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

teness

-0.73

andro

-0.73

bones

-0.72

MpServer

-0.70

ife

-0.69

SON

-0.68

ness

-0.67

ieft

-0.66

bone

-0.66

 Monster

-0.64

POSITIVE LOGITS

 shutter

1.01

âĸ¬

0.89

vation

0.79

CLASSIFIED

0.76

âĸ¬âĸ¬

0.74

bugs

0.71

itized

0.70

opsis

0.70

 Seym

0.68

 âĸº

0.65

Activations Density 0.022%

references to spatial locations or events

references to actions involving shutting or closing something

No Comments

No Known Activations