INDEX

Explanations

reports and mentions within text

oai_token-act-pair · gpt-3.5-turbo

repetitive mentions of the word "Reports."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 straw

-0.76

stroke

-0.72

 helm

-0.70

 collar

-0.66

 emotion

-0.66

 wink

-0.65

 wedge

-0.64

fit

-0.64

hug

-0.64

 kiss

-0.63

POSITIVE LOGITS

 Reports

3.75

Reports

2.81

reports

2.01

 Reporting

1.58

 Reported

1.57

 Sources

1.36

 Estimates

1.33

 Investigations

1.32

 reports

1.30

 Journals

1.28

Activations Density 0.012%

reports and mentions within text

repetitive mentions of the word "Reports."

No Comments

No Known Activations