INDEX

Explanations

words ending in '-ed' that are related to actions or descriptions

oai_token-act-pair · gpt-3.5-turbo

words related to hedging or uncertainty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

é¾įå

-0.71

 STOR

-0.61

VICE

-0.58

 Somalia

-0.56

DEF

-0.55

Vie

-0.54

BuyableInstoreAndOnline

-0.54

Sno

-0.54

 tatt

-0.54

 Wolves

-0.53

POSITIVE LOGITS

tons

0.96

ding

0.95

apeake

0.92

rals

0.88

rina

0.85

ochond

0.84

uling

0.82

rical

0.81

iva

0.81

uled

0.80

Activations Density 0.039%

words ending in '-ed' that are related to actions or descriptions

words related to hedging or uncertainty

No Comments

No Known Activations