INDEX

Explanations

phrases indicating difficulty or challenge

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

afety

-0.73

erity

-0.67

Phi

-0.66

calling

-0.64

plates

-0.63

 Presidential

-0.62

 Votes

-0.61

 Anthem

-0.61

 Stories

-0.61

 Brav

-0.61

POSITIVE LOGITS

 achieve

1.09

 quantify

1.06

 distinguish

1.04

 navigate

1.03

 reconcile

1.03

 imagine

1.03

 conceive

1.02

 attain

1.02

 obtain

1.02

 maintain

0.99

Activations Density 0.710%

phrases indicating difficulty or challenge

No Comments

No Known Activations