INDEX

Explanations

phrases indicating an increasing trend or comparison over time

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate increasing prevalence or quantity in various contexts

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

gee

-0.76

ORS

-0.73

MAT

-0.71

pload

-0.70

pper

-0.70

omics

-0.68

poke

-0.66

minimum

-0.66

assis

-0.66

minus

-0.66

POSITIVE LOGITS

 importantly

0.76

 invitations

0.72

 occurrences

0.72

 frustrated

0.71

 intrusive

0.70

 alike

0.68

 worrisome

0.67

 instances

0.66

 sophisticated

0.66

 gull

0.66

Activations Density 0.052%

phrases indicating an increasing trend or comparison over time

phrases that indicate increasing prevalence or quantity in various contexts

No Comments

No Known Activations