INDEX

Explanations

phrases related to medical treatments or interventions

oai_token-act-pair · gpt-3.5-turbo

references to various forms of medical treatment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

usercontent

-0.72

alsa

-0.72

allas

-0.71

umbnail

-0.70

Desktop

-0.69

estamp

-0.69

 Oops

-0.68

alias

-0.68

unes

-0.66

ulhu

-0.65

POSITIVE LOGITS

 treatment

3.76

treatment

2.99

 Treatment

2.93

 treatments

2.68

reatment

1.97

 therapy

1.91

treated

1.78

 treated

1.78

 treat

1.78

 treating

1.77

Activations Density 0.008%

phrases related to medical treatments or interventions

references to various forms of medical treatment

No Comments

No Known Activations