INDEX

Explanations

adjectives related to being attracted or compelled towards something

oai_token-act-pair · gpt-3.5-turbo

the word "drawn" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hal

-0.71

mma

-0.71

ican

-0.69

 VIDEOS

-0.69

duc

-0.67

lly

-0.67

eda

-0.65

ocker

-0.65

ership

-0.64

ateral

-0.63

POSITIVE LOGITS

 drawn

1.18

 Drawn

0.94

 drawing

0.81

 Drawing

0.78

drawn

0.77

GGGGGGGG

0.77

entious

0.73

LIB

0.70

 irresist

0.68

volt

0.68

Activations Density 0.007%

adjectives related to being attracted or compelled towards something

the word "drawn" in various contexts

No Comments

No Known Activations