INDEX

Explanations

instances of "tight" or concepts related to strong control or close-knit relationships

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

udi

-0.81

issance

-0.76

hex

-0.74

 è£ıç

-0.74

agne

-0.73

aurus

-0.72

umer

-0.71

oca

-0.70

;;;;;;;;;;;;

-0.69

opoulos

-0.68

POSITIVE LOGITS

 tight

1.00

 tighter

0.96

nesses

0.92

 toler

0.91

 tightly

0.90

 knit

0.89

 waist

0.88

 confinement

0.88

hold

0.87

neck

0.85

Activations Density 1.195%

instances of "tight" or concepts related to strong control or close-knit relationships

No Comments

No Known Activations