INDEX

Explanations

words related to destruction or physical collapse events

oai_token-act-pair · gpt-3.5-turbo

words related to structural failures or disasters

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

BIT

-0.75

thing

-0.72

Trend

-0.72

 Demand

-0.72

alty

-0.71

way

-0.70

inen

-0.70

ttle

-0.69

esthetic

-0.68

atre

-0.68

POSITIVE LOGITS

 subur

0.79

 corpses

0.75

 cabbage

0.74

ocument

0.74

©¶æ¥µ

0.74

 carc

0.71

 exting

0.70

 Leban

0.68

 Fukushima

0.68

 debris

0.68

Activations Density 0.111%

words related to destruction or physical collapse events

words related to structural failures or disasters

No Comments

No Known Activations