INDEX

Explanations

adjectives or phrases indicating importance, key points, or lessons learned

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

uild

-0.68

renheit

-0.63

rams

-0.61

netflix

-0.61

 Pastebin

-0.60

azard

-0.60

ð

-0.59

ancies

-0.59

liv

-0.59

blast

-0.59

POSITIVE LOGITS

 takeaway

1.00

 distinguishing

0.92

 downside

0.89

 concern

0.87

 concerns

0.82

 difference

0.81

 drawback

0.81

 thing

0.80

 reason

0.79

 caveat

0.75

Activations Density 2.407%

adjectives or phrases indicating importance, key points, or lessons learned

No Comments

No Known Activations