INDEX

Explanations

references to sports games where a tie occurred

oai_token-act-pair · gpt-3.5-turbo

references to competitive outcomes, particularly tying scores or standings in sports contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

âĵĺ

-0.79

resources

-0.79

isma

-0.75

rians

-0.74

illing

-0.73

zos

-0.71

xit

-0.68

 heterogeneity

-0.68

mining

-0.65

microsoft

-0.65

POSITIVE LOGITS

 streak

0.87

 championship

0.83

 Dwight

0.82

 teammate

0.81

MVP

0.80

 semifinals

0.79

 opponents

0.78

 playoff

0.77

 unbeaten

0.75

 opponent

0.75

Activations Density 0.105%

references to sports games where a tie occurred

references to competitive outcomes, particularly tying scores or standings in sports contexts

No Comments

No Known Activations