INDEX

Explanations

words related to a physical store or shop

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific store

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

SHIP

-0.73

vernment

-0.73

ersen

-0.72

theless

-0.69

vous

-0.68

param

-0.68

inx

-0.67

 Caucus

-0.65

ansson

-0.63

emonic

-0.62

POSITIVE LOGITS

 shelves

1.13

houses

0.98

 clerk

0.94

keepers

0.93

store

0.93

front

0.93

keeper

0.92

 store

0.92

 clerks

0.91

 stores

0.91

Activations Density 0.037%

words related to a physical store or shop

mentions of a specific store

No Comments

No Known Activations