INDEX

Explanations

references to work environments and activities

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

gall

-0.74

ocl

-0.64

alg

-0.63

egu

-0.63

icer

-0.61

Registered

-0.60

hest

-0.60

Redd

-0.60

hao

-0.58

ilde

-0.56

POSITIVE LOGITS

ago

0.96

 afterward

0.93

 afterwards

0.83

 thereafter

0.82

until

0.82

 uninterrupted

0.82

 periods

0.81

 until

0.77

 longer

0.76

iatus

0.73

Activations Density 0.535%

references to work environments and activities

No Comments

No Known Activations