INDEX

Explanations

ranges of values or quantities

oai_token-act-pair · gpt-3.5-turbo

terms that denote a continuous spectrum or variation

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

verse

-0.69

jiang

-0.68

mit

-0.64

orc

-0.63

 syndrome

-0.62

oliath

-0.61

elta

-0.61

æµ

-0.61

 root

-0.60

 reason

-0.60

POSITIVE LOGITS

ogyn

0.82

finder

0.79

range

0.78

ranging

0.78

 Emin

0.74

 ranging

0.73

 ranges

0.71

anges

0.70

 Luxem

0.68

Ĥª

0.67

Activations Density 0.017%

ranges of values or quantities

terms that denote a continuous spectrum or variation

No Comments

No Known Activations