INDEX

Explanations

programming-related terms related to function calls and features

oai_token-act-pair · gpt-3.5-turbo

programming structures and conditions in code

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

 Shuttle

-0.77

lus

-0.76

tnc

-0.73

 Hilton

-0.68

 Morales

-0.68

 Fiat

-0.65

 Peoples

-0.63

quin

-0.62

leck

-0.61

 Fairfax

-0.61

POSITIVE LOGITS

Required

0.78

*/

0.71

################

0.71

api

0.70

================================

0.70

################################

0.69

 json

0.67

Filename

0.67

 output

0.66

Activations Density 0.080%

programming-related terms related to function calls and features

programming structures and conditions in code

No Comments

No Known Activations