INDEX

Explanations

words related to explosives and bomb-related contexts

oai_token-act-pair · gpt-3.5-turbo

references to explosives or bomb-related terminology

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

laus

-0.83

DAY

-0.74

 Inquis

-0.67

dit

-0.66

}}}

-0.63

SEA

-0.63

 Quadro

-0.62

pring

-0.61

ç¥ŀ

-0.61

olars

-0.61

POSITIVE LOGITS

arding

1.29

arded

1.27

 detonated

1.08

 deton

1.05

 disposal

1.03

astically

1.00

astic

0.99

blast

0.90

 blast

0.88

proof

0.88

Activations Density 0.043%

words related to explosives and bomb-related contexts

references to explosives or bomb-related terminology

No Comments

No Known Activations