INDEX

Explanations

references to prominent figures or characters, often in sports or entertainment 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

gently

-0.71

NEC

-0.69

CTR

-0.69

 Giul

-0.59

ilty

-0.59

rency

-0.58

 forth

-0.57

issance

-0.56

forth

-0.56

 abdom

-0.56

POSITIVE LOGITS

otta

0.77

oshenko

0.76

awk

0.74

lamm

0.70

emouth

0.70

arkin

0.70

acca

0.69

etz

0.68

arella

0.68

udi

0.67

Activations Density 0.019%

references to prominent figures or characters, often in sports or entertainment contexts

No Comments

No Known Activations