INDEX

Explanations

mentions of specific historic events or pop culture 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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

yet

-0.78

externalActionCode

-0.70

ENGTH

-0.68

IDA

-0.68

ajor

-0.68

 Sanctuary

-0.67

sett

-0.67

hd

-0.66

eem

-0.65

ells

-0.63

POSITIVE LOGITS

entimes

1.30

pmwiki

1.13

 resorted

0.83

 overlooked

0.83

 involve

0.83

times

0.83

 occasions

0.82

 referred

0.81

 occurrence

0.81

 forget

0.79

Activations Density 5.806%

mentions of specific historic events or pop culture references

No Comments

No Known Activations