INDEX

Explanations

mentions of enjoyable food items or rewards

oai_token-act-pair · gpt-3.5-turbo

references to treats or special foods

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

 constitu

-0.69

 autonomous

-0.67

 moot

-0.65

ova

-0.64

 Karin

-0.62

seless

-0.62

dc

-0.59

 condem

-0.59

 Citiz

-0.59

 Dani

-0.59

POSITIVE LOGITS

ises

1.12

ties

0.95

ise

0.95

nels

0.94

itionally

0.90

pieces

0.90

piece

0.90

orial

0.87

terson

0.85

ery

0.85

Activations Density 0.024%

mentions of enjoyable food items or rewards

references to treats or special foods

No Comments

No Known Activations