INDEX

Explanations

specific details or information

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

ammers

-0.69

inals

-0.67

bey

-0.66

ove

-0.64

nder

-0.64

ereo

-0.63

owship

-0.63

artney

-0.60

 Ferdinand

-0.60

upiter

-0.60

POSITIVE LOGITS

 details

1.22

 detail

1.14

displayText

0.94

 detailing

0.94

 thereof

0.90

 specifics

0.90

 wording

0.88

 glean

0.88

 redacted

0.87

 outline

0.87

Activations Density 0.999%

specific details or information

No Comments

No Known Activations