INDEX

Explanations

mentions of strongholds or protected areas

oai_token-act-pair · gpt-3.5-turbo

negative or derogatory terms, particularly "bast" and its variations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odcast

-0.69

 complication

-0.64

MRI

-0.63

viously

-0.61

ERA

-0.60

ysis

-0.60

 instruction

-0.60

orsi

-0.60

rose

-0.57

 Speaker

-0.57

POSITIVE LOGITS

ados

0.98

mouth

0.90

hips

0.77

chart

0.75

alde

0.72

 sectors

0.72

hip

0.71

udo

0.70

ãĥİ

0.69

burg

0.69

Activations Density 0.017%

mentions of strongholds or protected areas

negative or derogatory terms, particularly "bast" and its variations

No Comments

No Known Activations