INDEX

Explanations

references to specific locations or organizations, particularly with the term "Stoke" and "Bray"

oai_token-act-pair · gpt-3.5-turbo

names of locations related to sports teams or cities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

guyen

-0.70

 skelet

-0.66

HCR

-0.66

å¥³

-0.64

 Russ

-0.64

inarily

-0.64

ADE

-0.64

schild

-0.64

umbn

-0.63

trak

-0.61

POSITIVE LOGITS

rament

0.91

 Stoke

0.90

burgh

0.84

borough

0.82

emouth

0.77

wick

0.77

 City

0.75

ring

0.70

icket

0.70

 Hots

0.69

Activations Density 0.006%

references to specific locations or organizations, particularly with the term "Stoke" and "Bray"

names of locations related to sports teams or cities

No Comments

No Known Activations