INDEX

Explanations

instances of change or transformation within a context

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

ccoli

-0.71

isson

-0.70

ilyn

-0.69

ogg

-0.68

ONY

-0.68

staking

-0.67

osher

-0.66

lehem

-0.66

onda

-0.65

audi

-0.65

POSITIVE LOGITS

 fortunes

1.08

 trajectory

1.00

 outlook

0.92

 tack

0.92

 behaviour

0.92

 course

0.90

 existing

0.89

 perception

0.89

 behavior

0.88

 wording

0.86

Activations Density 1.518%

instances of change or transformation within a context

No Comments

No Known Activations