INDEX

Explanations

implicitly stated information or biases

oai_token-act-pair · gpt-3.5-turbo

references to implicit or tacit knowledge and biases

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ionage

-0.83

gard

-0.83

ishers

-0.78

yrs

-0.72

 Fighters

-0.69

hner

-0.68

rar

-0.68

THING

-0.67

affles

-0.67

reth

-0.67

POSITIVE LOGITS

 implicit

1.11

 tacit

1.00

 guiActiveUn

1.00

 implicitly

0.96

 assumption

0.86

 inference

0.78

osion

0.78

 imply

0.76

 complicity

0.75

verse

0.75

Activations Density 0.018%

implicitly stated information or biases

references to implicit or tacit knowledge and biases

No Comments

No Known Activations