INDEX

Explanations

instances of words/phrases denoting sufficiency or importance

oai_token-act-pair · gpt-3.5-turbo

phrases indicating capability or sufficiency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĥĨ

-0.79

chwitz

-0.68

 Mehran

-0.67

updated

-0.64

Â®,

-0.63

Products

-0.63

ãĥ©ãĥ³

-0.63

 Glob

-0.62

cycl

-0.61

uned

-0.61

POSITIVE LOGITS

 warrant

1.59

 merit

1.31

 justify

1.25

 outweigh

1.12

 qualify

1.11

 withstand

1.05

 satisfy

1.04

 deserve

1.03

 bother

1.00

 suffice

1.00

Activations Density 0.150%

instances of words/phrases denoting sufficiency or importance

phrases indicating capability or sufficiency

No Comments

No Known Activations