INDEX

Explanations

references to legal or political issues

oai_token-act-pair · gpt-3.5-turbo

phrases related to potential developments and challenges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 tremend

-0.90

âĢĲ

-0.81

 inval

-0.81

 enthusi

-0.76

 ensu

-0.76

âĹ¼

-0.76

————————

-0.73

ilst

-0.72

 scrut

-0.72

————————————————

-0.71

POSITIVE LOGITS

Washington

1.31

CNN

1.06

 Replay

0.87

JUST

0.84

Photos

0.76

CNN

0.75

GOP

0.74

Authorities

0.73

Adds

0.72

Obama

0.72

Activations Density 0.046%

references to legal or political issues

phrases related to potential developments and challenges

No Comments

No Known Activations