INDEX

Explanations

concepts related to catastrophic events or potential disasters

oai_token-act-pair · gpt-3.5-turbo

references to dramatic changes or overwhelming situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

andom

-0.64

Basically

-0.58

phia

-0.57

information

-0.56

 everybody

-0.55

Information

-0.55

 whats

-0.55

ãĤ£

-0.55

 agreement

-0.54

 doesnt

-0.54

POSITIVE LOGITS

 examples

0.77

Such

0.73

such

0.63

ĸļ

0.63

 Such

0.62

 Examples

0.60

say

0.60

 attest

0.59

 Fukushima

0.59

 commonplace

0.58

Activations Density 1.495%

concepts related to catastrophic events or potential disasters

references to dramatic changes or overwhelming situations

No Comments

No Known Activations