INDEX

Explanations

phrases with commatas

oai_token-act-pair · gpt-3.5-turbo

overall percentages related to public opinions or demographics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

minster

-0.84

iously

-0.81

RL

-0.71

aciously

-0.70

MAX

-0.66

inar

-0.66

rou

-0.66

Laughs

-0.66

bor

-0.64

ãĥ¥

-0.64

POSITIVE LOGITS

 however

1.26

 meanwhile

1.00

 moreover

0.93

 though

0.90

 according

0.86

 although

0.83

 excluding

0.79

 therefore

0.77

 suffice

0.75

 analysts

0.74

Activations Density 0.411%

phrases with commatas

overall percentages related to public opinions or demographics

No Comments

No Known Activations