INDEX

Explanations

mentions of beginning or end markers ('Re', 'St', 'Cl', 'Ch') in the text

oai_token-act-pair · gpt-3.5-turbo

terms related to recognition and representation in various contexts

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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.69

è¦ļéĨĴ

-0.67

ĵĺ

-0.67

ĸļ

-0.67

 respectively

-0.65

[&

-0.63

MODE

-0.62

 initials

-0.62

zona

-0.59

 thereto

-0.59

POSITIVE LOGITS

vertising

1.09

roximately

1.09

jamin

1.01

resa

0.98

withstanding

0.95

icester

0.94

anie

0.94

cohol

0.94

xiety

0.89

odore

0.87

Activations Density 0.220%

mentions of beginning or end markers ('Re', 'St', 'Cl', 'Ch') in the text

terms related to recognition and representation in various contexts

No Comments

No Known Activations