INDEX

Explanations

sentences related to counting or measuring

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ieties

-0.66

Oracle

-0.65

 conditioning

-0.64

 obser

-0.63

 Harbor

-0.62

 Bree

-0.62

 waivers

-0.61

insula

-0.61

 Centauri

-0.60

 habit

-0.59

POSITIVE LOGITS

enance

1.78

downs

0.89

esses

0.87

rified

0.86

ess

0.82

ries

0.77

icates

0.77

ensen

0.76

icated

0.74

eenth

0.74

Activations Density 0.629%

sentences related to counting or measuring

No Comments

No Known Activations