INDEX

Explanations

references to something being legendary

oai_token-act-pair · gpt-3.5-turbo

instances of the word "legendary" 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 @ 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

elling

-0.88

CDC

-0.86

arettes

-0.84

chens

-0.80

vere

-0.80

getting

-0.79

ells

-0.78

eling

-0.77

yg

-0.76

itten

-0.76

POSITIVE LOGITS

 antiquity

0.85

 proportions

0.79

 adventurer

0.78

 Creature

0.78

theless

0.77

 ancestor

0.76

 achievement

0.75

 relic

0.75

 tales

0.75

 treasures

0.74

Activations Density 0.020%

references to something being legendary

instances of the word "legendary" and its variations

No Comments

No Known Activations