INDEX

Explanations

words related to the term "For" and possibly titles or concepts associated with it

oai_token-act-pair · gpt-3.5-turbo

high-frequency words related to geographical locations or institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

bour

-0.65

onite

-0.65

 Cosponsors

-0.63

habi

-0.63

wagen

-0.62

hari

-0.61

ved

-0.61

ouri

-0.60

opted

-0.60

Bar

-0.59

POSITIVE LOGITS

 sake

1.36

 purposes

0.99

 foreseeable

0.75

 reasons

0.73

ummies

0.70

 upcoming

0.70

 upkeep

0.68

erity

0.68

 clarification

0.65

ankind

0.65

Activations Density 0.913%

words related to the term "For" and possibly titles or concepts associated with it

high-frequency words related to geographical locations or institutions

No Comments

No Known Activations