INDEX

Explanations

words ending in '-ing' that denote ongoing actions or states

oai_token-act-pair · gpt-3.5-turbo

present participles and gerunds

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

ensable

-0.76

xual

-0.70

ONSORED

-0.64

ãĥĲ

-0.62

Ł

-0.56

ENN

-0.55

è£ħ

-0.54

encia

-0.54

ãĥķãĤ©

-0.53

enment

-0.53

POSITIVE LOGITS

redients

0.88

aukee

0.83

mouth

0.79

bery

0.78

tips

0.77

addle

0.77

tons

0.76

tails

0.76

ulate

0.73

bing

0.72

Activations Density 0.085%

words ending in '-ing' that denote ongoing actions or states

present participles and gerunds

No Comments

No Known Activations