INDEX

Explanations

words and phrases related to toxins or poisonous substances

oai_token-act-pair · gpt-3.5-turbo

references to poison and poisoning

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

noon

-0.86

irc

-0.86

ann

-0.80

blance

-0.75

ONT

-0.72

Struct

-0.72

held

-0.72

ear

-0.71

AMA

-0.71

dar

-0.69

POSITIVE LOGITS

 poison

1.26

 poisoning

1.26

 poisoned

1.14

 poisonous

1.10

 poisons

1.00

 dart

0.99

 arsenic

0.94

 antidote

0.93

 darts

0.89

 Poison

0.84

Activations Density 0.009%

words and phrases related to toxins or poisonous substances

references to poison and poisoning

No Comments

No Known Activations