INDEX

Explanations

instances of the word "let"

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

natureconservancy

-0.75

existent

-0.61

 holiest

-0.56

omething

-0.55

 veins

-0.55

Interstitial

-0.55

ombat

-0.55

 verte

-0.53

 Âł Âł Âł Âł

-0.53

 wielded

-0.53

POSITIVE LOGITS

itia

1.04

icia

0.94

ting

0.93

tering

0.92

tered

0.91

arations

0.85

hetically

0.82

us

0.79

ugal

0.75

ters

0.74

Activations Density 0.021%

instances of the word "let"

No Comments

No Known Activations