INDEX

Explanations

proper nouns related to specific individuals or places

oai_token-act-pair · gpt-3.5-turbo

words or parts of words that contain the sequence "ph."

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

 footing

-0.65

ciating

-0.63

 segregation

-0.60

 Collider

-0.59

enthal

-0.57

 Accord

-0.57

 anch

-0.57

 srfAttach

-0.57

 Documents

-0.57

 Hardcore

-0.57

POSITIVE LOGITS

adelphia

1.30

ebus

1.09

aceutical

0.90

ogram

0.84

thora

0.84

ongs

0.81

ases

0.80

ascript

0.79

dit

0.78

Ļ

0.77

Activations Density 0.070%

proper nouns related to specific individuals or places

words or parts of words that contain the sequence "ph."

No Comments

No Known Activations