INDEX

Explanations

descriptive terms or actions related to disruptive or damaging transformation processes

oai_token-act-pair · gpt-3.5-turbo

terms related to degradation or decline

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

OWS

-0.83

Reviewer

-0.82

razil

-0.80

glers

-0.75

STER

-0.71

 Holmes

-0.71

aneers

-0.69

ONY

-0.66

amia

-0.66

intendent

-0.65

POSITIVE LOGITS

ync

0.98

 resil

0.92

ktop

0.91

irable

0.88

perate

0.87

 embr

0.86

erve

0.85

semb

0.83

erving

0.80

iple

0.79

Activations Density 0.005%

descriptive terms or actions related to disruptive or damaging transformation processes

terms related to degradation or decline

No Comments

No Known Activations