INDEX

Explanations

phrases related to physical actions or confrontations involving force

oai_token-act-pair · gpt-3.5-turbo

variations of the word "hang" in different 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

duc

-0.80

ibaba

-0.73

XP

-0.72

plin

-0.71

xia

-0.71

pees

-0.71

ç«

-0.70

veland

-0.68

ledged

-0.68

ISO

-0.68

POSITIVE LOGITS

erness

0.92

ishly

0.85

wise

0.82

anged

0.80

anging

0.78

lass

0.75

hetic

0.74

lies

0.73

enegger

0.72

 Wonders

0.72

Activations Density 0.020%

phrases related to physical actions or confrontations involving force

variations of the word "hang" in different contexts

No Comments

No Known Activations