INDEX

Explanations

phrases related to specific locations or places

oai_token-act-pair · gpt-3.5-turbo

locations and geographical references

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

 positives

-0.47

CHAT

-0.44

apego

-0.43

rehensive

-0.43

onite

-0.42

 carbohyd

-0.41

bre

-0.40

isites

-0.39

PDATE

-0.39

ophy

-0.38

POSITIVE LOGITS

 Indies

0.61

é¾įå¥ĳå£«

0.56

 vicinity

0.50

 guise

0.50

 circles

0.49

 Netherlands

0.48

rooms

0.48

 Tanzania

0.48

 Fiji

0.47

 Royale

0.47

Activations Density 1.489%

phrases related to specific locations or places

locations and geographical references

No Comments

No Known Activations