INDEX

Explanations

repeated mentions of the word "any"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

 elig

-0.69

isable

-0.69

 dilig

-0.66

 fung

-0.66

PDATE

-0.60

 unemploy

-0.60

ITNESS

-0.60

reditary

-0.60

 leptin

-0.59

 surv

-0.58

POSITIVE LOGITS

where

1.03

one

0.92

ika

0.92

body

0.87

emi

0.83

uan

0.83

elle

0.81

heter

0.81

thood

0.80

THING

0.80

Activations Density 0.005%

repeated mentions of the word "any"

No Comments

No Known Activations