INDEX

Explanations

responses or reactions to various prompts or cues

oai_token-act-pair · gpt-3.5-turbo

variations of the word "respond" and its related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Rwanda

-0.82

 geographically

-0.72

BALL

-0.67

 dividing

-0.66

 onto

-0.63

 Sabha

-0.62

 warp

-0.62

 twists

-0.61

piece

-0.61

GGGGGGGG

-0.61

POSITIVE LOGITS

onding

1.30

ibilities

1.21

onds

1.16

iration

1.13

ents

1.11

awn

1.11

itable

1.06

ublic

1.06

encies

1.04

onder

1.03

Activations Density 0.061%

responses or reactions to various prompts or cues

variations of the word "respond" and its related forms

No Comments

No Known Activations