INDEX

Explanations

mentions of regions or locations in India

oai_token-act-pair · gpt-3.5-turbo

references to specific geographical locations, particularly places in Jammu and Kashmir, and Newfoundland

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hiba

-0.95

arin

-0.81

TPP

-0.78

ucl

-0.76

hod

-0.76

icles

-0.76

yss

-0.74

illard

-0.74

aic

-0.74

oping

-0.73

POSITIVE LOGITS

wark

0.95

matter

0.88

uitous

0.83

ãĤ¡

0.78

orkshire

0.75

itionally

0.73

ãĥĳ

0.72

sburgh

0.70

uit

0.69

ãĤ¼ãĤ¦ãĤ¹

0.69

Activations Density 0.025%

mentions of regions or locations in India

references to specific geographical locations, particularly places in Jammu and Kashmir, and Newfoundland

No Comments

No Known Activations