INDEX

Explanations

scientific article references

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

ħĭ

-0.84

Ļ½

-0.65

ipel

-0.64

ĸļ

-0.62

 chin

-0.62

hiba

-0.61

 unrealistic

-0.60

itia

-0.58

olate

-0.58

edi

-0.58

POSITIVE LOGITS

0.89

0.87

Downloadha

0.74

pn

0.73

bryce

0.73

zip

0.71

google

0.69

jpg

0.68

html

0.68

Activations Density 0.034%

scientific article references

No Comments

No Known Activations