INDEX

Explanations

names containing the sequence "ih"

oai_token-act-pair · gpt-3.5-turbo

the presence of individuals or characters in a narrative context

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

yards

-0.70

 Predator

-0.66

âĸ¬

-0.65

mint

-0.64

 Chips

-0.63

 Pengu

-0.61

 Colossus

-0.60

 Monstrous

-0.59

ãĥ¯

-0.59

 Firefly

-0.59

POSITIVE LOGITS

onen

1.15

ollow

1.04

ype

1.00

ih

0.99

ouston

0.96

ansen

0.96

azard

0.96

oney

0.95

ackers

0.93

awk

0.93

Activations Density 0.010%

names containing the sequence "ih"

the presence of individuals or characters in a narrative context

No Comments

No Known Activations