INDEX

Explanations

words ending in '-hers'

oai_token-act-pair · gpt-3.5-turbo

terms related to ownership or possession

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

RAW

-0.64

gged

-0.60

ements

-0.60

lif

-0.59

nces

-0.58

 repayment

-0.58

ANC

-0.58

clusive

-0.57

 Picard

-0.57

Gall

-0.56

POSITIVE LOGITS

mith

1.35

chool

1.10

hers

0.91

hare

0.84

ervatives

0.83

itage

0.83

enium

0.83

creen

0.81

ands

0.81

uggest

0.81

Activations Density 0.012%

words ending in '-hers'

terms related to ownership or possession

No Comments

No Known Activations