INDEX

Explanations

mentions of the country "India" or things related to it

oai_token-act-pair · gpt-3.5-turbo

references to India and associated entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

nect

-0.79

mble

-0.77

cer

-0.75

icle

-0.75

cies

-0.74

 Kessler

-0.73

ording

-0.73

tnc

-0.72

rd

-0.71

fman

-0.69

POSITIVE LOGITS

apolis

1.07

 Sharma

0.97

 Pradesh

0.95

ru

0.93

 Punjab

0.92

Rao

0.91

 Kumar

0.91

 Gupta

0.90

Pv

0.89

 cricket

0.86

Activations Density 0.057%

mentions of the country "India" or things related to it

references to India and associated entities

No Comments

No Known Activations