INDEX

Explanations

references to health-related actions or treatments

oai_token-act-pair · gpt-3.5-turbo

references to dietary changes or health-related nutrition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

###

-0.69

requisites

-0.61

Enlarge

-0.60

 dives

-0.60

 reserved

-0.59

 honorable

-0.58

......

-0.58

DL

-0.58

flight

-0.58

.........

-0.57

POSITIVE LOGITS

 Sabha

0.79

ptin

0.74

0.72

oresc

0.71

orie

0.70

ocrates

0.69

awaru

0.68

orescent

0.67

borgh

0.67

acters

0.66

Activations Density 0.000%

references to health-related actions or treatments

references to dietary changes or health-related nutrition

No Comments

No Known Activations