INDEX

Explanations

references to wooded areas or forests

oai_token-act-pair · gpt-3.5-turbo

references related to wooded areas and wilderness

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

activation

-0.69

 contagious

-0.68

Published

-0.67

OPLE

-0.66

andum

-0.65

interstitial

-0.64

Companies

-0.64

ICT

-0.62

 cler

-0.61

inel

-0.61

POSITIVE LOGITS

 woods

1.16

thur

0.98

hed

0.96

 woodland

0.92

leigh

0.86

lake

0.81

woods

0.81

geoning

0.79

mere

0.79

 creek

0.78

Activations Density 0.008%

references to wooded areas or forests

references related to wooded areas and wilderness

No Comments

No Known Activations