INDEX

Explanations

negatively opinionated terms related to product performance and failures

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

quartered

-0.68

agen

-0.68

 Travels

-0.66

 Â®

-0.64

 Koran

-0.63

 Nobel

-0.62

 awaits

-0.61

 Guant

-0.60

ikh

-0.60

 Shiite

-0.56

POSITIVE LOGITS

 inexper

0.74

 outwe

0.74

 beforehand

0.71

Overall

0.69

 anyways

0.69

needed

0.65

Was

0.64

 synergy

0.63

 minded

0.63

ranged

0.62

Activations Density 0.762%

negatively opinionated terms related to product performance and failures

No Comments

No Known Activations