INDEX

Explanations

pieces of text indicating someone expressing their opinion or sharing information

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a need or action to be taken

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

hur

-0.66

river

-0.62

Motion

-0.62

ById

-0.61

Ĭ±

-0.61

 Miko

-0.60

ahime

-0.59

 Lines

-0.59

 largeDownload

-0.58

âĢ²

-0.58

POSITIVE LOGITS

 overcome

0.92

 prove

0.92

 accomplish

0.88

 prepare

0.84

 survive

0.84

 achieve

0.83

 improve

0.83

 contend

0.83

 learn

0.82

 protect

0.81

Activations Density 0.080%

pieces of text indicating someone expressing their opinion or sharing information

phrases that indicate a need or action to be taken

No Comments

No Known Activations