INDEX

Explanations

actions related to interaction and communication

oai_token-act-pair · gpt-3.5-turbo

phrases related to interpersonal conflict and communication issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

eatures

-0.76

rafted

-0.70

senal

-0.69

everal

-0.66

inently

-0.65

 convergence

-0.64

accompan

-0.63

etheus

-0.62

uilt

-0.61

uably

-0.61

POSITIVE LOGITS

 anymore

1.93

 whatsoever

1.19

nor

1.16

 anyways

1.15

 unless

1.13

 anything

1.09

anything

1.05

 haha

1.04

 anyway

1.04

.</

1.02

Activations Density 0.358%

actions related to interaction and communication

phrases related to interpersonal conflict and communication issues

No Comments

No Known Activations