INDEX

Explanations

numbers written as words

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letters "EN" in various contexts

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

 Cind

-0.68

 Shoals

-0.66

Reference

-0.65

ts

-0.64

 Miracle

-0.64

tie

-0.61

axe

-0.61

 Heights

-0.59

mson

-0.59

mie

-0.58

POSITIVE LOGITS

ET

1.09

TERN

1.08

vironment

1.07

ERAL

1.06

ISH

1.04

NER

0.99

CLA

0.98

AME

0.96

ABLE

0.95

OVA

0.94

Activations Density 0.012%

numbers written as words

occurrences of the letters "EN" in various contexts

No Comments

No Known Activations