INDEX

Explanations

words related to locations or specific places, specifically gymnasiums

oai_token-act-pair · gpt-3.5-turbo

references to gymnasiums

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

 snipp

-0.69

 outgoing

-0.67

 Mush

-0.65

 Kore

-0.65

 possessions

-0.64

 seizures

-0.63

 REPL

-0.62

 Sunshine

-0.62

 bitters

-0.61

theless

-0.60

POSITIVE LOGITS

zu

0.95

ppo

0.92

nt

0.84

acus

0.80

xx

0.79

ci

0.78

ournals

0.78

vid

0.77

zona

0.76

idence

0.75

Activations Density 0.000%

words related to locations or specific places, specifically gymnasiums

references to gymnasiums

No Comments

No Known Activations