INDEX

Explanations

expressions related to forecasting or prediction

oai_token-act-pair · gpt-3.5-turbo

references to time, planning, and social welfare issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 condem

-0.67

 denotes

-0.51

FML

-0.48

 acknow

-0.47

 begs

-0.47

 transpired

-0.47

 warr

-0.47

 pleading

-0.47

 retracted

-0.46

 announces

-0.46

POSITIVE LOGITS

by

0.87

by

0.87

through

0.86

rather

0.84

without

0.81

 rather

0.80

By

0.76

By

0.75

via

0.74

—

0.73

Activations Density 0.710%

expressions related to forecasting or prediction

references to time, planning, and social welfare issues

No Comments

No Known Activations