INDEX

Explanations

mention of specific institutions and organizations

oai_token-act-pair · gpt-3.5-turbo

nouns associated with entities, classifications, or systems

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

asonable

-0.72

ogenous

-0.65

opping

-0.65

istical

-0.64

fitting

-0.64

cia

-0.64

urate

-0.62

uran

-0.61

otta

-0.61

typically

-0.60

POSITIVE LOGITS

Ħ¢

0.67

LTD

0.66

pora

0.66

 consortium

0.64

 detachment

0.63

 Explorer

0.62

naire

0.60

 Netherlands

0.59

 Sweep

0.59

 Telescope

0.58

Activations Density 0.322%

mention of specific institutions and organizations

nouns associated with entities, classifications, or systems

No Comments

No Known Activations