INDEX

Explanations

names and titles related to individuals

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

 Palest

-0.73

 Adin

-0.68

eele

-0.67

 Kessler

-0.65

 fortun

-0.63

å£

-0.59

-"

-0.57

intern

-0.56

comed

-0.55

ãĥķãĤ©

-0.55

POSITIVE LOGITS

ONSORED

0.83

vernight

0.74

//[

0.69

endif

0.68

 Journals

0.66

 caption

0.65

APTER

0.64

itionally

0.61

--------------------------------------------------------

0.61

<-

0.60

Activations Density 1.235%

names and titles related to individuals

No Comments

No Known Activations