INDEX

Explanations

terms related to hockey

oai_token-act-pair · gpt-3.5-turbo

references to hockey

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ibility

-1.05

terday

-0.84

ible

-0.82

iary

-0.77

ages

-0.74

igm

-0.73

iment

-0.72

antis

-0.71

utter

-0.70

AGES

-0.70

POSITIVE LOGITS

 rink

1.18

 puck

1.00

 hockey

0.99

 Hockey

0.93

 Club

0.85

 league

0.80

 League

0.78

 goalie

0.77

 club

0.77

 playoff

0.76

Activations Density 0.010%

terms related to hockey

references to hockey

No Comments

No Known Activations