INDEX

Explanations

references to nuclear weapons and arms control issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

MpServer

-0.88

 Homeless

-0.81

ĸļ

-0.81

idav

-0.80

 "$:/

-0.80

Ort

-0.79

 Gallagher

-0.77

idas

-0.75

Ļ½

-0.75

knit

-0.73

POSITIVE LOGITS

 warheads

1.22

 submarines

1.19

 plutonium

1.16

 submarine

1.14

 nuclear

1.14

 detonated

1.13

 deton

1.12

iferation

1.10

 missile

1.09

utonium

1.08

Activations Density 0.144%

references to nuclear weapons and arms control issues

No Comments

No Known Activations