INDEX

Explanations

references to explosive materials

oai_token-act-pair · gpt-3.5-turbo

terminology related to explosives and explosive materials

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

Sea

-0.77

early

-0.72

SEA

-0.72

Son

-0.72

DAY

-0.72

 Presidents

-0.71

aird

-0.71

erest

-0.71

 Attention

-0.70

come

-0.70

POSITIVE LOGITS

 explosives

1.17

 detonated

1.10

 deton

1.06

 barric

0.94

abwe

0.92

 grenades

0.88

avorite

0.88

querque

0.85

 disposal

0.83

 breaching

0.82

Activations Density 0.008%

references to explosive materials

terminology related to explosives and explosive materials

No Comments

No Known Activations