INDEX

Explanations

words indicating intensity or extremity in descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

usterity

-0.71

ragon

-0.68

romeda

-0.65

ossal

-0.64

terday

-0.63

rador

-0.63

 Locked

-0.62

olation

-0.60

heast

-0.60

unct

-0.59

POSITIVE LOGITS

 saying

1.02

say

0.93

SAY

0.82

 Saying

0.78

Say

0.72

 said

0.70

 stating

0.69

 says

0.68

 argue

0.67

FFER

0.65

Activations Density 0.403%

words indicating intensity or extremity in descriptions

No Comments

No Known Activations