INDEX

Explanations

references to specific sports events, players, and actions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

\":

-0.69

ISTORY

-0.67

PDATED

-0.67

 ILCS

-0.65

ãĥ£

-0.62

 renaissance

-0.57

)=

-0.56

phas

-0.54

mble

-0.54

iencies

-0.54

POSITIVE LOGITS

 whom

0.90

via

0.82

existent

0.77

 while

0.76

 whenever

0.72

 inappropriately

0.70

omever

0.70

 because

0.69

 beforehand

0.69

 ASAP

0.68

Activations Density 5.198%

references to specific sports events, players, and actions

No Comments

No Known Activations