INDEX

Explanations

mentions or references to reports or news articles

oai_token-act-pair · gpt-3.5-turbo

occurrences 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

enfranch

-0.87

vantage

-0.75

arching

-0.67

breaker

-0.67

pired

-0.67

uzzle

-0.66

igmatic

-0.66

rial

-0.62

 resonate

-0.61

 hinge

-0.61

POSITIVE LOGITS

 reports

3.60

reports

2.94

 Reports

2.56

Reports

2.53

 report

2.15

 reported

2.03

 reporting

1.96

 Report

1.89

report

1.89

Report

1.72

Activations Density 0.020%

mentions or references to reports or news articles

occurrences of the word "reports."

No Comments

No Known Activations