INDEX

Explanations

mentions of news agencies and photographers

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

gone

-0.70

gran

-0.69

 sediment

-0.67

 reinvest

-0.66

 tranquil

-0.66

hal

-0.65

sed

-0.65

within

-0.61

 chilling

-0.61

izers

-0.60

POSITIVE LOGITS

sburg

1.11

 Images

1.08

PLIED

0.81

sburgh

0.79

Images

0.77

aukee

0.77

 IMAGES

0.77

 Features

0.76

PLE

0.74

PLA

0.73

Activations Density 1.113%

mentions of news agencies and photographers

No Comments

No Known Activations