INDEX

Explanations

information related to studies or research findings on specific topics

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

inventoryQuantity

-0.85

othe

-0.75

 goodbye

-0.75

 farewell

-0.70

CHAPTER

-0.69

bye

-0.69

 trad

-0.67

EngineDebug

-0.67

enance

-0.67

 tant

-0.66

POSITIVE LOGITS

RESULTS

1.44

Results

1.11

 Results

1.10

results

0.92

ULTS

0.90

 Participants

0.87

sample

0.86

 Subjects

0.86

 Surprisingly

0.86

 findings

0.85

Activations Density 0.343%

information related to studies or research findings on specific topics

No Comments

No Known Activations