INDEX

Explanations

names of individuals and references to their actions or roles

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

ailability

-0.66

 citiz

-0.59

 conflic

-0.58

icable

-0.58

BIL

-0.57

nah

-0.56

 Seym

-0.56

 Kardash

-0.55

 Celsius

-0.55

etheless

-0.54

POSITIVE LOGITS

zzi

1.00

ieri

0.88

igham

0.84

Jr

0.78

's

0.73

xus

0.72

 Presents

0.70

zzo

0.69

inelli

0.69

Sr

0.69

Activations Density 0.137%

names of individuals and references to their actions or roles

No Comments

No Known Activations