INDEX

Explanations

phrases related to names

oai_token-act-pair · gpt-3.5-turbo

words related to performances and comedic acts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 tert

-0.69

emetery

-0.69

 Skydragon

-0.66

 Overse

-0.65

 visitation

-0.64

 quarters

-0.63

 Airl

-0.63

 "$:/

-0.63

 lieu

-0.62

CFR

-0.62

POSITIVE LOGITS

puff

0.92

pel

0.90

auga

0.87

rill

0.84

opol

0.82

inki

0.81

ifiers

0.81

iland

0.80

abi

0.78

ify

0.78

Activations Density 0.059%

phrases related to names

words related to performances and comedic acts

No Comments

No Known Activations