INDEX

Explanations

keywords related to notable titles or achievements, particularly in a contextual or official capacity

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

));

-0.71

 decl

-0.64

 waivers

-0.64

 reliably

-0.61

 uncertain

-0.61

lining

-0.58

 glance

-0.58

 captcha

-0.56

xp

-0.55

 Weston

-0.55

POSITIVE LOGITS

estine

1.06

foundland

0.94

odore

0.92

anmar

0.89

cellence

0.86

etheus

0.85

Zero

0.84

tainment

0.82

ageddon

0.82

agate

0.82

Activations Density 0.122%

keywords related to notable titles or achievements, particularly in a contextual or official capacity

No Comments

No Known Activations