INDEX

Explanations

references to descent or decline, particularly in contexts related to people or situations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

acea

-0.85

omaly

-0.63

archives

-0.62

keeping

-0.62

 reass

-0.62

ako

-0.62

stuff

-0.61

bara

-0.59

iso

-0.58

isations

-0.58

POSITIVE LOGITS

 descended

0.84

ENTS

0.81

ents

0.80

 descend

0.80

 upon

0.78

 into

0.76

lycer

0.68

ient

0.67

LECT

0.67

river

0.66

Activations Density 0.012%

references to descent or decline, particularly in contexts related to people or situations

No Comments

No Known Activations