INDEX

Explanations

phrases related to operating tasks or machinery

oai_token-act-pair · gpt-3.5-turbo

references to operators in a computational or procedural context

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

ilan

-0.72

ioxide

-0.69

 Agric

-0.65

 Emin

-0.65

loe

-0.62

urd

-0.61

imm

-0.61

uran

-0.61

ances

-0.60

enser

-0.59

POSITIVE LOGITS

otomy

0.94

 operator

0.89

 Operator

0.83

operator

0.79

 operators

0.77

querade

0.74

 licenses

0.72

operated

0.70

eering

0.70

 licences

0.66

Activations Density 0.016%

phrases related to operating tasks or machinery

references to operators in a computational or procedural context

No Comments

No Known Activations