INDEX

Explanations

references to conversations or discussions

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

aples

-0.71

 Scotia

-0.66

uilt

-0.65

 Allies

-0.63

Constructed

-0.63

 Soldier

-0.61

undai

-0.60

rette

-0.59

 unforeseen

-0.59

bered

-0.58

POSITIVE LOGITS

ative

0.94

ership

0.93

radio

0.82

about

0.82

ers

0.81

tion

0.80

osphere

0.80

 about

0.78

uced

0.77

ularity

0.77

Activations Density 0.021%

references to conversations or discussions

No Comments

No Known Activations