INDEX

Explanations

player names from different sports or activities

oai_token-act-pair · gpt-3.5-turbo

specific proper nouns and identifiers in a text

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

cember

-0.72

ĸļ

-0.68

STD

-0.67

aucuses

-0.64

WAY

-0.63

uddin

-0.60

ipeg

-0.59

dor

-0.59

bia

-0.58

ftime

-0.58

POSITIVE LOGITS

sic

0.77

tein

0.66

ilts

0.63

auld

0.62

 Expand

0.62

chio

0.61

 extraord

0.59

*/(

0.59

iggs

0.58

 Tactics

0.57

Activations Density 0.892%

player names from different sports or activities

specific proper nouns and identifiers in a text

No Comments

No Known Activations