INDEX

Explanations

words related to logic, arguments, and reasoning

oai_token-act-pair · gpt-3.5-turbo

concepts related to reason, principles, and opportunities 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

 convol

-0.67

orb

-0.66

FY

-0.65

 blat

-0.63

IRC

-0.62

ordable

-0.61

nep

-0.61

ammy

-0.61

ynski

-0.60

ean

-0.59

POSITIVE LOGITS

lessly

0.94

fulness

0.94

lessness

0.82

fully

0.81

abl

0.79

nesia

0.76

ful

0.72

icism

0.71

making

0.70

 Incarn

0.69

Activations Density 0.503%

words related to logic, arguments, and reasoning

concepts related to reason, principles, and opportunities in various contexts

No Comments

No Known Activations