INDEX

Explanations

phrases that have been explicitly explained or detailed

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "outlined."

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

airs

-0.67

awar

-0.66

oping

-0.65

alle

-0.65

cess

-0.65

otion

-0.65

asions

-0.65

isol

-0.64

agent

-0.64

cgi

-0.63

POSITIVE LOGITS

 outlines

1.10

 outline

1.06

 outlined

1.04

 outlining

1.03

 sket

0.85

 showc

0.85

ĸļ

0.82

escription

0.81

ocument

0.80

WER

0.79

Activations Density 0.010%

phrases that have been explicitly explained or detailed

occurrences of the word "outlined."

No Comments

No Known Activations