INDEX

Explanations

expressions of uncertainty and recommendations

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

Bridge

-0.67

ilts

-0.63

legal

-0.60

Applic

-0.60

 Moves

-0.59

 merger

-0.57

 rendition

-0.57

undown

-0.56

 Returns

-0.56

Op

-0.55

POSITIVE LOGITS

 yourselves

1.41

 yourself

1.39

 your

0.89

 Yourself

0.88

Tube

0.88

 YOUR

0.81

your

0.77

agine

0.73

tube

0.72

 Tube

0.72

Activations Density 1.513%

expressions of uncertainty and recommendations

No Comments

No Known Activations