INDEX

Explanations

technical terms related to architecture and design

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

rious

-0.62

 aggrav

-0.57

itatively

-0.55

llah

-0.55

 Haram

-0.54

 nause

-0.54

pard

-0.53

keley

-0.52

 positively

-0.52

 affirmative

-0.52

POSITIVE LOGITS

0.88

}}

0.78

odox

0.68

 âĢº

0.68

senal

0.67

)|

0.67

 Reports

0.67

Politics

0.66

utterstock

0.65

 Files

0.64

Activations Density 0.210%

technical terms related to architecture and design

No Comments

No Known Activations