INDEX

Explanations

countries or geopolitical entities

oai_token-act-pair · gpt-3.5-turbo

specific place names and associated subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

insula

-0.60

 VIDEOS

-0.60

ortment

-0.59

 Flavoring

-0.59

çĭ

-0.58

romeda

-0.58

lihood

-0.57

robe

-0.57

 Nanto

-0.57

UNE

-0.57

POSITIVE LOGITS

ivable

0.81

 incarn

0.80

arily

0.78

isible

0.77

rozen

0.77

iful

0.77

urated

0.76

apped

0.75

centric

0.75

urized

0.74

Activations Density 0.481%

countries or geopolitical entities

specific place names and associated subjects

No Comments

No Known Activations