INDEX

Explanations

texts related to laws, regulations, and political news

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

usercontent

-0.73

 unless

-0.70

!.

-0.68

".

-0.67

mpire

-0.66

marine

-0.66

.(

-0.66

unless

-0.64

$.

-0.63

'.

-0.63

POSITIVE LOGITS

 foregoing

0.73

math

0.61

 disclaim

0.58

 hindsight

0.57

¿½

0.56

 Hits

0.55

 HuffPost

0.55

cknowled

0.54

 careful

0.52

 retiring

0.51

Activations Density 2.464%

texts related to laws, regulations, and political news

No Comments

No Known Activations