INDEX

Explanations

mentions of specific years and seasons

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

esse

-0.74

-0.65

ocker

-0.63

 lately

-0.62

overe

-0.61

had

-0.60

gotten

-0.59

 recent

-0.59

 recently

-0.59

effic

-0.59

POSITIVE LOGITS

 onwards

0.89

utical

0.72

 attRot

0.66

................................

0.65

lins

0.65

rica

0.65

iatus

0.64

 unless

0.64

lease

0.64

soon

0.63

Activations Density 0.036%

mentions of specific years and seasons

No Comments

No Known Activations