INDEX

Explanations

instances of the word "from" and phrases indicating choices or origins

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

mination

-0.84

matter

-0.84

mouth

-0.84

riter

-0.81

scient

-0.77

visor

-0.76

notice

-0.76

haw

-0.74

hai

-0.74

ended

-0.74

POSITIVE LOGITS

 preset

0.90

 among

0.89

 assorted

0.87

 amongst

0.86

 various

0.81

 whichever

0.80

 options

0.79

 afar

0.78

 dozens

0.78

 available

0.77

Activations Density 0.023%

instances of the word "from" and phrases indicating choices or origins

No Comments

No Known Activations