INDEX

Explanations

instances of the phrase "up" in various contexts

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

iquid

-0.61

eers

-0.61

 Monaco

-0.61

 Goldstein

-0.61

DCS

-0.60

 Cust

-0.60

 Abyss

-0.60

MAT

-0.58

 specialization

-0.57

 Guerrero

-0.57

POSITIVE LOGITS

raised

1.10

rights

0.98

 stairs

0.88

stairs

0.83

stretched

0.80

rompt

0.80

 asleep

0.78

river

0.77

skirts

0.77

 beside

0.77

Activations Density 0.021%

instances of the phrase "up" in various contexts

No Comments

No Known Activations