INDEX

Explanations

phrases related to responding to various statements or events

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "resp" or its variations related to response or reactions

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

66666666

-0.78

 whispers

-0.75

gling

-0.70

mere

-0.70

 ISBN

-0.69

 twists

-0.68

 directions

-0.67

 devils

-0.67

JFK

-0.65

POSITIVE LOGITS

Resp

1.18

onding

1.10

awn

1.07

onds

1.00

rha

0.91

 Resp

0.90

onder

0.90

responsive

0.90

ixel

0.89

TPPStreamerBot

0.87

Activations Density 0.006%

phrases related to responding to various statements or events

occurrences of the word "resp" or its variations related to response or reactions

No Comments

No Known Activations