INDEX

Explanations

instances of the word "from" indicating sources or origins in the text

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

itivity

-0.82

ifix

-0.74

hyde

-0.71

seek

-0.71

enary

-0.70

hops

-0.70

finals

-0.70

stakes

-0.69

hop

-0.68

ratulations

-0.66

POSITIVE LOGITS

 afar

1.25

 whence

0.96

 thence

0.92

 inception

0.82

 obscurity

0.77

0.74

 standpoint

0.73

 perspective

0.73

0.72

 conception

0.72

Activations Density 0.029%

instances of the word "from" indicating sources or origins in the text

No Comments

No Known Activations