INDEX

Explanations

questions or statements involving knowledge or information

oai_token-act-pair · gpt-3.5-turbo

references to guessing and knowledge-based questioning

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Contents

-0.76

egu

-0.73

UTC

-0.71

Dialogue

-0.69

imm

-0.68

itals

-0.68

egal

-0.68

enge

-0.68

roman

-0.67

²¾

-0.67

POSITIVE LOGITS

 Oops

0.80

 Pledge

0.74

 Want

0.74

Won

0.74

 Bike

0.71

 Spice

0.70

 naughty

0.70

 kidding

0.70

 Bastard

0.69

 Didn

0.68

Activations Density 0.319%

questions or statements involving knowledge or information

references to guessing and knowledge-based questioning

No Comments

No Known Activations