INDEX

Explanations

adjectives ending in 'ant' and 'ent'

oai_token-act-pair · gpt-3.5-turbo

words that indicate ongoing or persistent conditions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

©¶æ

-0.94

illet

-0.85

mberg

-0.80

mbuds

-0.79

adr

-0.78

soDeliveryDate

-0.76

herty

-0.76

creen

-0.75

dule

-0.74

ernels

-0.74

POSITIVE LOGITS

ial

0.95

aneous

0.91

ty

0.90

ness

0.88

ially

0.88

iate

0.77

 minded

0.76

iary

0.75

ly

0.74

iated

0.73

Activations Density 0.043%

adjectives ending in 'ant' and 'ent'

words that indicate ongoing or persistent conditions

No Comments

No Known Activations