INDEX

Explanations

specific references to the term "Lehman Brothers" and various forms of the word "man" in the context of financial discussions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

externalActionCode

-0.86

ILE

-0.83

hered

-0.76

urally

-0.70

 appropri

-0.70

 biom

-0.69

amazon

-0.69

HCR

-0.69

population

-0.67

 UNESCO

-0.66

POSITIVE LOGITS

 Brothers

1.28

 Bros

1.06

 Sachs

1.04

 Holdings

1.00

 Horowitz

0.85

 Corporation

0.83

 bankruptcy

0.83

 Investments

0.82

 Investors

0.82

 Companies

0.78

Activations Density 0.016%

specific references to the term "Lehman Brothers" and various forms of the word "man" in the context of financial discussions

No Comments

No Known Activations