INDEX

Explanations

phrases related to observations or perceptions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "seeing"

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

prime

-0.81

 whistle

-0.80

cult

-0.80

cussion

-0.69

chip

-0.69

si

-0.68

sett

-0.67

locked

-0.67

cover

-0.66

trap

-0.64

POSITIVE LOGITS

seeing

0.80

 seeing

0.79

wcsstore

0.79

-+-+

0.77

enance

0.75

agame

0.74

ometimes

0.74

angelo

0.71

ablishment

0.71

 mete

0.70

Activations Density 0.014%

phrases related to observations or perceptions

instances of the word "seeing"

No Comments

No Known Activations