INDEX

Explanations

phrases describing how something looks or appears

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate perceptions or appearances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 coun

-0.64

 railing

-0.62

 Surv

-0.61

cember

-0.61

ItemTracker

-0.60

 Ways

-0.59

PsyNetMessage

-0.58

urus

-0.58

 Storage

-0.57

arius

-0.57

POSITIVE LOGITS

vind

0.89

eless

0.81

 stronger

0.78

 smarter

0.76

 invincible

0.74

operative

0.73

keeper

0.71

 unbeat

0.70

llular

0.69

overe

0.69

Activations Density 0.082%

phrases describing how something looks or appears

phrases that indicate perceptions or appearances

No Comments

No Known Activations