INDEX

Explanations

references to physical or metaphorical openings

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

rior

-0.74

CTV

-0.71

VICE

-0.67

grade

-0.66

Bey

-0.64

ãĥīãĥ©ãĤ´ãĥ³

-0.64

orum

-0.63

rag

-0.62

 constitu

-0.61

iable

-0.60

POSITIVE LOGITS

 doors

1.10

 Doors

1.09

 portals

0.89

 loopholes

0.83

up

0.82

ource

0.81

 wounds

0.79

doors

0.77

 curtains

0.77

 reopen

0.76

Activations Density 0.396%

references to physical or metaphorical openings

No Comments

No Known Activations