INDEX

Explanations

mentions of the word "positivity" or words related to a positive outlook

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "pos" and related variations, indicating a focus on positioning or status

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

loo

-0.90

HAEL

-0.76

 Hearts

-0.70

clad

-0.68

 Clarkson

-0.67

stall

-0.66

 oats

-0.66

ARD

-0.64

OPS

-0.64

 Continental

-0.63

POSITIVE LOGITS

itional

1.30

itions

1.27

idon

1.17

itory

1.14

itivity

1.10

itor

1.09

itives

1.07

pos

1.04

itionally

1.01

sel

0.98

Activations Density 0.021%

mentions of the word "positivity" or words related to a positive outlook

occurrences of the term "pos" and related variations, indicating a focus on positioning or status

No Comments

No Known Activations