INDEX

Explanations

the name "Jake" at different levels of relevance

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Jake."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

iary

-0.80

iated

-0.76

 conduc

-0.74

amera

-0.70

iating

-0.67

 scoreboard

-0.64

iaries

-0.64

oppable

-0.63

IFIED

-0.63

journal

-0.63

POSITIVE LOGITS

glers

0.95

Gy

0.93

unin

0.93

ansas

0.92

cki

0.81

xtap

0.81

 McGee

0.81

 Skywalker

0.77

Bone

0.76

 Jake

0.75

Activations Density 0.029%

the name "Jake" at different levels of relevance

mentions of the name "Jake."

No Comments

No Known Activations