INDEX

Explanations

numerical formats with specific patterns

oai_token-act-pair · gpt-3.5-turbo

patterns of scoring or performance metrics in sports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Feder

-0.65

 Gors

-0.64

Jaw

-0.61

 Quartz

-0.61

 treat

-0.61

 Sund

-0.61

 horrified

-0.61

 Examiner

-0.60

esa

-0.60

 Reader

-0.59

POSITIVE LOGITS

WAR

0.85

win

0.81

ranked

0.81

0.80

RGB

0.79

EXP

0.79

neutral

0.77

Activations Density 0.044%

numerical formats with specific patterns

patterns of scoring or performance metrics in sports

No Comments

No Known Activations