INDEX

Explanations

words related to the action of pressing or the term "press"

oai_token-act-pair · gpt-3.5-turbo

references to "press" in various contexts, indicating a focus on media or publication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ooters

-0.67

 racial

-0.66

 Aber

-0.65

 aver

-0.63

 Median

-0.63

 deductions

-0.61

 nightmare

-0.61

Obj

-0.60

 overtime

-0.59

 drunken

-0.59

POSITIVE LOGITS

press

4.71

Press

2.44

PRESS

2.23

 press

1.79

pressed

1.79

 Press

1.79

 PRESS

1.73

 presses

1.72

 pressing

1.12

print

1.06

Activations Density 0.014%

words related to the action of pressing or the term "press"

references to "press" in various contexts, indicating a focus on media or publication

No Comments

No Known Activations