INDEX

Explanations

phrases related to being affected or injured by something or someone

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate actions or events influenced by something or someone

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ASE

-0.83

eva

-0.77

inea

-0.76

rest

-0.74

ãĥĥãĥĪ

-0.73

cue

-0.72

iture

-0.71

LO

-0.71

yssey

-0.70

ourage

-0.70

POSITIVE LOGITS

products

0.95

 virtue

0.95

 illness

0.82

 sheer

0.81

 gunfire

0.77

 nature

0.76

 lightning

0.75

 mobs

0.75

 Hurricane

0.74

 inertia

0.73

Activations Density 0.111%

phrases related to being affected or injured by something or someone

phrases that indicate actions or events influenced by something or someone

No Comments

No Known Activations