INDEX

Explanations

phrases related to feeling ill or unwell

oai_token-act-pair · gpt-3.5-turbo

terms related to energy and lethality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ipedia

-0.87

 Britann

-0.68

edom

-0.67

ulates

-0.66

Meh

-0.66

reddits

-0.65

uler

-0.65

 hypoc

-0.64

ocratic

-0.64

olicy

-0.63

POSITIVE LOGITS

wcsstore

0.84

sie

0.84

ãĤī

0.80

 Paso

0.80

zie

0.80

elligence

0.80

swick

0.78

ILA

0.76

LV

0.75

apo

0.75

Activations Density 0.027%

phrases related to feeling ill or unwell

terms related to energy and lethality

No Comments

No Known Activations