INDEX

Explanations

the word "fun" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to enjoyment and fun activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ographers

-0.73

cially

-0.72

cale

-0.70

oak

-0.69

Report

-0.67

ACP

-0.67

ographer

-0.66

faced

-0.66

åº

-0.65

rise

-0.65

POSITIVE LOGITS

fun

3.73

fun

2.04

Fun

1.86

FUN

1.78

Fun

1.68

 enjoyable

1.61

 entertaining

1.42

 enjoyment

1.33

 hilar

1.30

joy

1.30

Activations Density 0.016%

the word "fun" in various contexts

references to enjoyment and fun activities

No Comments

No Known Activations